NAP — Nanopore sequencing-derived Amplicon Pipeline

By Luke B. Jones

---

Requirements

1. Conda
2. Hardware: A PC capable of handling large datasets (or alternatively, sufficient time).
   Recommended: >20 GB RAM and a CPU with >4 cores.

---

How to set up

# Clone the repository and set up NAP, refresh bashrc to ensure PATH is working
git clone https://github.com/Luke-B-Jones/NAP.git
cd ./NAP
./build.sh
source ~/.bashrc

# The conda environment should be named 'nap_env'.
conda env create -f environment.yaml

# If 'nap update-database' fails, check that your ~/.bashrc  
# has been correctly updated to include 'nap' as an alias.

# If you intend to use primers that are not default, skip this final step and see the User Manual (Parts 1 and 2).
conda activate nap_env
nap update-database SILVA_138.2_SSU_NR99

---

How to use

Always ensure your project’s input data is stored in ./raw_data before running any commands,
unless you choose to modify the wrapper to suit your own directory structure.

# Example: Setting up contamination controls  
# (assumes ./raw_data/*13*.fasta corresponds to B1 - Blank 1)

nap pipe 13 B1 14 B2 15 B3
nap decon ./B1/B1*SPECIES-LEVEL.tsv ./B2/B2*SPECIES-LEVEL.tsv ./B3/B3*SPECIES-LEVEL.tsv

Once decontamination is enabled, proceed to data analysis.
You can confirm this by checking config.sh:
if the variable blank_read_count has content and blank_active="1", you are ready to proceed.

# Example: Processing samples  
# (assumes ./raw_data/*13*.fasta corresponds to S1 - Sample 1)

nap pipe 3 S1 2 S2 1 S3

---

Interpretation

1. Pipe score: Ranges from 0–100, calculated using Phred quality and read count.
   Values >90 are considered high-quality outputs.

2. Due to the nature of nanopore sequencing-based amplicons,
   low-abundance artefacts may be present in any run scoring <90.

3. Assuming ~20% of reads are >Q30, we recommend a raw input of
   >500,000 reads per sample.

---

User Manual

1. Database setup

The classification database is set up via the following process:

1. The user downloads the desired database.

2. scripts/update-database performs conserved processing steps, and
   defers filtering to scripts/mammalian_microbiome_inclusion.sh, which by default removes:

   • uncultured entries
   • metagenome-derived entries
   • unclassified entries
   • chlorophyll-derived sequences
   • unlikely eukaryotes (for microbiome research)

3. Remaining reference reads are trimmed to isolate amplified regions,
   using the primer configuration specified in subconfigs/AMP* (as set in config.sh).

Users wishing to use a different database should:

• examine how the database is annotated
• update the awk commands in scripts/mammalian_microbiome_inclusion.sh (lines 21 and 29) accordingly
• consider the classification conventions used by their chosen database

To configure a new primer set, users should:

1. Create a new subconfig:
   subconfigs/AMP_${primer_name}.sh
2. Update config.sh:
   export amplicon_pre_set="${primer_name}"
3. Re-run database setup.

---

2. Custom amplicons

Custom amplicons are added by creating a new file in:

subconfigs/AMP*.sh

(where * is the primer name), and activating it in config.sh via:

amplicon_pre_set=""

Users should follow the template and guidance in:

subconfigs/AMP_template.sh

---

3. Using non-barcoded raw data

Inside the nap wrapper (around line 99), the script extracts the path of a barcoded dataset
based on the numeric identifier supplied on the command line.

Example:

nap pipe 14 S1

This instructs the pipeline to use:

./raw_data/*barcode14.fastq

as the raw input for sample S1.

Users with differently named raw data may safely modify this part of the wrapper.

---

Thanks to

• Morgan Cockrill (MSc, University of Bath) — Improved robustness of pipe modules and QC section
• Josephine Ilott (MSc, University of Bath) — Authored the decontamination Python module